INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. 1/18

Liquefaction of Saturated Sandy Soils

Liquéfaction de sols sableux saturés

by V. A. F lorin , Professor, Doctor of Technical Sciences, Associate Member of the USSR Academy of Sciences, Institute of Mechanics of the USSR Academy of Sciences, and P. L. I vanov , Docent, Candidate of Technical Sciences, Leningrad Polytechnic Institute.

Summary Sommaire The authors give results obtained from theoretical and Le présent rapport contient quelques résultats des études experimental research on the liquefaction of saturated sandy théoriques et expérimentales des phénomènes de liquéfaction soils undertaken by them, in the Laboratory of Soil Mechanics de sols sableux saturés qui ont été effectuées aux Laboratoires of the Institute of Mechanics of the U.S.S.R. Academy of Sciences de Mécanique des sols de l’institut de Mécanique de l’Académie and in the Leningrad Polytechnic Institute. des Sciences de l’U.R.S.S. et de l’institut Polytechnique de Laboratory investigations have been carried out on various Leningrad. kinds of disturbances affecting saturated sandy soils and causing Les auteurs décrivent des études de Laboratoire de la liqué­ their liquefaction. The authors also give the results of studies faction de sols sableux soumis à diverses influences. on the grain-size composition and density of sands, shape of Ils exposent l’influence sur les phénomènes de liquéfaction de particles, entrapped gas content, initial stress condition, intensity la granulométrie et de la densité apparente des sables, de la forme and character of disturbances causing a collapse of the structure des grains, de la teneur en gaz occlus, de l’état initial de contraintes of the soil. Theoretical relationships are given which connect du sol, de l’intensité et du caractère des influences provoquant the process of consolidation in the liquefied soil with the time la rupture de la structure du sol. during which the sand is in a liquid state. These relationships Ils proposent des relations théoriques, relatives au processus largely determine the probable movements of both structures de consolidation des particules d’un sol liquéfié permettant en and liquified masses of soil. particulier, le calcul de la durée pendant laquelle le sable se Finally, the « method of standard blasting » worked out by trouve à l’état de liquéfaction. Ces relations déterminent en V. A. Florin and widely used in the Soviet Union is recommended grande partie les mouvements possibles d’ouvrages et de masses as being suitable for field investigations of the phenomenon de sol liquéfié. reviewed. Pour les études « in situ » du phénomène de liquéfaction, on propose d’employer « la méthode d’explosions uniformes » mise au point par V. A. Florin. La méthode est largement employée en U.R.S.S.

The liquefaction of saturated sandy soils may be described different areas in homogeneous soil, regardless of the static as the mechanical breakdown of the structure of the sand homogeneity of the total volume of soil, the stress condition as well as by physical and chemical processes which tend to is different for each individual particle. Different shape and occur in soils, and in particular, by thixotropy. (This is the size of isolated particles account for a different num ber of property possessed by some gels of becoming fluid when contacts between them and different stresses. As a result, even stirred, and of returning to the jelly state when stirring ceases). insignificant disturbances like filtration will cause not only This paper discusses only those cases of soil liquefaction complete but also partial collapse of the structure of the sand which are observed in pure sands and which are of a m echa­ due to displacements of individual unstable particles of sand nical nature. The phenomenon of complete or partial loss of subjected to minimum stress. By averaging the variable the bearing capacity of a soil, also the transition of the sand quantity of the displaced sand particles, the authors obtain into the state of flow due to collapse of the structure and dis­ “the degree of break-down”, of the structure of the sand and as placements of grains of saturated sand followed by the form a­ a result “the degree of liquefaction”, of the sand. T he degree of tion of a new, denser sediment and by the decrease o f the breakdown of the structure, or liquefaction may be expressed porosity of the sand, should be described as liquefaction of as a percentage of the breakdown of contacts, displacem ent sandy soil. of all particles and complete liquefaction of the soil. Liquefaction consists in breaking down the structure, and The effect of liquefaction on the strength and stability o f liquefying the soil mass itself with subsequent consolidation. structures is evaluated not only by its amount of liquefaction, Thus the conditions necessary for liquefaction are : the but also by its character. The possible movements o f structures collapse of the structure, with the possibility of sand consolida­ are largely influenced by the time during which the sand tion and either partial or complete saturation of the sand in a liquid state and by the viscosity of the liquefied mass of with water. the soil. Because of the great variety of factors causing the collapse The time of liquefaction depends upon the thickness of of a sand structure, the authors propose that the criterion of the liquefied stratum of the sand, the permeability o f the collapse of the soil structure should be not the “critical void sand, variation in the volume of voids in the process of ratio” or the density of the sand, but “critical” values of the consolidation, intensity of draining, location of the drainage intensity of dynamic disturbance, stress condition of the soil system in the body of a structure and the duration of dynamic or weight of surcharge, and hydraulic gradient of w ater flow load causing collapse of the sand structure. through it. For several years the authors carried out numerous labora­ Sands may pass into a state of either complete or partial tory and large-scale field investigations on several large hydro­ liquefaction. Considering individual particles of the sand or electric developments. Under laboratory conditions, tests

107 were performed on special impact-and-vibrating tables and The initial distribution of head in a stratum of com pletely in flumes. Tests were also carried out under practical cond­ liquefied sand may be expressed by the following equation itions on sands which were placed both naturally and hydrau- (Fig. 2) : lically. The main features of sand transformation into the liquefied h = ^?(.S — x) + x , (i) state are the decrease of its porosity and a temporary increase T w of poor water pressure. Under these conditions, measurem ents where : have been made of changes in the porosity of the sand by Yj, = y' + y ,„ — unit weight of completely liquefied soil, using deeply sited reference points and an electrical m ethod Y' — unit weight of suspended sand, to measure porosity, based on the relationship between Y,„ — unit weight of water. the specific electrical conductivity of the soil and that of the pore water. Simultaneously changes in the pore water pres­ U— 5 — I sures have been measured using special membrane-type inductive cells. In addition, observations were made on the velocity at which heavy bodies sank into the sand. In impulse tests, loose saturated sands were readily concer­ ted into a completely liquefied state over the entire depth of the stratum. The curves representing maximum excess hydraulic pressure were of triangle shape (Fig. 1 a) and the

Fig. 2 Distribution of heads in a stratum of completely liquefied sand. Répartition des charges dans la couche de sol totalement liquéfiée. The density of the sand changes, in general, at the contact 0 5 10 15 20 surfaces between the liquefied and non-liquefied sand stratum. excess poze pzessuie . cm. of w atez caCumn Composing the equation of water equilibrium in elem entary Fig. 1 Fluctuations of maximum excess water pressure due to boundary layer, and considering the conditions for seepage successive impact influences and consolidation of of the squeezed water through a medium, we obtain the follo­ a sand stratum in completely liquefied state. wing : Variations des surpressions maxima dans l’eau sous Yu l’effet de percussions successives et consolidation de (2) la souche sableuse totalement liquéfiée. K y ' 1 — « where : pressure corresponded to the weight of the waterlogged K — coefficient of seepage, stratum with suspended sand particles (y p <=» 2 gm/cm3). W ith n0 — initial porosity, the increase of sand density, the influence of stress condition n — porosity after sand consolidation. of the sand determined by the weight of overburden tends to be greater, and in the lower layers of the soil no liquefaction The curves representing heads and velocity of movem ent occurs. Because of the pressure of contact between sand p a r­ of the boundary of the liquefied soil obtained by equation (2) ticles, greater intensity of dynamic disturbance is required agree very well with the experimental values. to break down the structure of the sand than the structure In contrast to shocks, vibrations do not cause liquefaction of its upper layers, and as a result the curves representing over the whole stratum of the soil. The process proceeds in maximum excess pressures at subsequent impulses are of layers : the first vibrations liquefy the upper layer which trapezoidal shape (Fig. 1 a). carries a comparatively light load. This reduces the pressures The consolidation of sand in complete liquefaction does of overburden on the lower layers and as a result the latter not depend on the intensity of dynamic disturbance, and in pass into the liquid state with later vibrations. T he zone of the case of local liquefaction the increase of the intensity of liquefaction extends downwards. (Fig. 3a). As soon as the dynamic disturbance leads to expansion of the liquefied zone only but not to an increase in density of sand consol­ idation. As soon as the sand passes into a state of complete lique­ faction, the replacement of particles or sand consolidation takes place followed by squeezing the water out of the voids in the soil. The curves representing the distribution of excess water pressure for different times, are of trapezoidal shape (Fig. 1 b). The shape of these curves is evidence th at for each moment of time there is a boundary between the consol­ idated part of the sand and the sand in a liquid state. The process of particle replacement starts in the lower part of the excès? poze pzessuie, cm. of wate 7 coeumn sand stratum, the boundary between the consolidated and liquefied sand being moved in the direction of the surface of Fig. 3 Distribution of excess pressure in a sand stratum acted the soil. The upper layers of the soil remain liquefied m uch on by vibrations. longer, and in these layers there is much greater displacements Répartition des surpressions dans une couche de sable of the liquefied soil or maximum sinking of heavy bodies. soumise aux effets de la vibration. 108 entire stratum of sand is liquefied, the process of particle replacement or consolidation occurs, pressures in the pore water decrease and the boundary of the consolidated soil moves upwards. The supplementary consolidation which takes place in the consolidation portion of the soil is due to the action of conti­ nuous vibrations (Fig. 3 c). Many engineers believe that only fine sands may be lique­ fied. However, experience based on laboratory investigations has shown that all sufficiently loose, cohesionless soils of any grain-size may be liquefied, becauses this is only an inter­ mediate phase in the process of consolidation of cohesionless soils, acted on dynamically. Due to high permeability, the time during which the liquid state lasts is much less for coarse-grained soils than for fine-grained ones (Fig. 4). Since

30

ft.

j 20 1 numffez of impacts

« e Fig. 5 Influence of the intensity of draining surcharge on the period of time within which the sand remains liquid. o Influence de l’intensité de la surcharge de drainage sur 0,25 0,50 0,75 100 125 le temps pendant lequel la couche de sable se trouve mm, à l’état liquéfié. Fig. 4 Influence of grain-size composition of sand on the period of time within which the sand remains liquid. Influence de la granulométrie des sables sur le temps h 5 q. pendant lequel la couche de sable se trouve à l’état liquéfié.

the liquid state lasts for only a short time, the liquefied masses of soil have no time for displacements, so that there is prac­ tically no indication that the phenomenon of liquefaction occurs in coarse-grained soils. The initial stress condition for saturated sandy soil deter­ mined by the weight of the overburden or external applied load, reduces the possibility of sand liquefaction. Increase of loads demands considerable increase of the dynam ic dis­ turbance required to cause a collapse of the structure of the sand and to liquefy it. As a result of many field observations, the authors reached the conclusion that at depths ranging from about 10 to 15 metres below ground level, even very loose sand can hardly be liquefied. Surcharge with any Fig. 6 Distribution of heads in a stratum of liquefied sand material can be used as a method of reducing sand liquefaction. influenced by draining surcharge. The use of draining surcharge reduces not only the possibi­ Répartition des charges dans la couche de sable liquéfié lity of sand liquefaction but also, if liquefaction has already en présence de la surcharge de drainage. occured, the period during which it will continue. As the impulse tests have shown, even small draining surcharge where : will reduce the time of the liquid state tenfold. (Fig. 5). q 1 — n With a draining surcharge the process of secondary consol­ A = — K.-,— • ------; idation of soil goes on simultaneously both in upper and -y„, n0 — n lower parts of the liquefied sand stratum (Fig. 6). In the lower part of the stratum consolidation is due to the weight of B = —• -X [jf (1 _ „ ) _ Ko( 1 — «)]. particles. This consolidation may be evaluated by equation «o — " Y w (2). In the upper part of the stratum consolidation is due to From equations (2) and (3) we can determine the time the weight of the surcharge. Considering the conditions for during which the sand stratum will remain in a liquid state. seepage of the water squeezed out of the voids in the soil With the increase of stress due to the weight of the soil itself, water, we can obtain the following relationship which enables or to an externally applied load, the compaction of the sand us to locate the upper boundary of the liquefied stratum at considerably decreases with the result that the duration of different intervals of time : the liquid state will also be reduced. B Drains in the sand mass also provide the conditions for 1 + - (S — r2) ■ (S — r 2) rapid expulsion of water from the voids in the liquefied soil. A The time during which a 5-metre thickness of fine sand (3) remains liquefied is given by equation (2) and is 27 minutes, 109 with the draining surcharge of 5 kg/cm 2 — 2 minutes, and with the supplementary drainage at the base of the stratum — it ranges from 20 to 15 seconds. This reduction of the time during which the sand remains liquefied results in im proved stability both of liquefied soil and of structures built upon it. Shear strength of a completely liquefied sand is determ ined mainly by its viscosity. If the viscosity is sufficiently great, no Fig. 7 Settlement of the surface of the soil after blasting of three standard charges. significant displacements of the liquefied sand occur even at the longer time duration of the sand in a liquid state. Tassement de la surface de sol après trois explosions uniformes. With increase of density, the viscosity of the liquefied sand increase 20-30 times, the latter having been observed when the unit weight of the soil is considerably less than th at of 15-25 minutes. Within the limits of the upper non-saturated the soil in the extremely dense condition. part of the soil a net of concentric fissures will be form ed Laboratory and field investigations have shown that the due to settlement of the surface. The length of som e o f the presence of entrapped gas, the content of which may vary from fissures was 20-25 m. Heavy bodies tend to sink in a liquefied 8 to 20 per cent, greatly influences the process of liquefaction soil ; for example, steel reinforcement bars 50 mm in dia­ of the sand. The presence of entrapped gas considerably meter and 8-10 kg by weight sank as deep as 1-5-2-0 m. decreases the intensity and velocity of propagation (up to Blasting of standard charges in dense sands results in less 100 — 50 m/sec at 7 to 8 per cent of gas content) of dynamic settlement, and in several cases it does not occur at all. There disturbances and consequently decreases the zone of lique­ is no visible evidence of liquefaction and neither do heavy faction. With the increase of entrapped gas content the com ­ bodies sink into the liquefied soil. paction of the sand due to dynamic disturbance decreases, The main criteria for liquefaction of saturated sandy soils and loose sands with relatively small entrapped gas content in relation to standard blasting are the magnitude of average do not pass into a liquid state. settlement of the surface of the soil in a radius o f 5 metres The sands which are uniform according to their grain-size from the place of explosion and the ratio of settlement resulted composition with the same unit weight or porosity m ay have from three successive standard blastings at one place. Tests different densities due to different shape of particles. F or have shown that in case the average settlement in a radius example, the porosity of fine poorly rounded sand which has of 5 metres is less than 8-10 cm there is no need to provide a limited zone of liquefaction is 45 per cent, while the porosity measures against liquefaction. The greater the difference o f of sand with analogous grain-size composition but with well average settlement after three successive blastings, the looser rounded particles is 38 per cent. It may be advisable to use will be the sand and the greater is the possibility of liquefaction. the analogy method for evaluating the liquefaction o f sandy If the ratio between two successive blastings is greater than soil with caution, because this method is applicable only to 1:0,6 then there is danger of arising of an undesirable spreading the grain-size composition of the sand. of liquefaction. Thus, the laboratory tests have shown that in several The method of standard blasting was used to study the cases the phenomenon of liquefaction causes neither significant possibility of liquefaction of natural and hydraulically filled deformation nor failure of structures. Therefore, in the design soil on certain Russian rivers. of structures the assumption regarding the possibility of The results of some of the tests are presented in the Table, liquefaction can be made, provided that due allowance is which follows. made for possible harmful effects. Liquefaction which is not followed by spreading of the soil causes only a settlem ent of the surface of the soil, and the permissible value should Average be determined for each particular case depending on the Poro­ Ground Number Settle­ Radius o f settle­ structure and the purpose for which it is intended. It is n o te­ No. Soils sity water of ment in per level explo­ a radius ment worthy that most hydraulic-fill dams do not require special extension measures to be taken againt sand liquefaction. cent meters sions o f 5 m cm m Investigations also proved that liquefaction propagation of a “chain reaction” type does not exist in nature. If the surface of the soil is horizontal no liquefaction propagation is likely 1 Sand, fine, to occur. The local transformation of the sand into a liquid subrounded 44-45 0-5 1 22 25-30 state on the inclined surface of the soil may cause the loss of 2 — 43 10 1 17 20-22 — stability in the adjacent dry soil, followed by failure of this 3 41-42 0-3 1 16 16-18 —— 2 10 part of the soil. The necessary condition for such propagation 3 8 is a certain steepness of slope. 4 — 0-2 1 10 10-12 Research into the possibility of liquefying saturated sandy 2 8 soils under field conditions may be successfully performed by 3 5 using explosives for deep blasting. In order to evaluate the 5 —— 0-3 1 1 8 results of investigations and to compare blastings for different 6 Sand, fine, 38 0-2 1 5 10 soil conditions, tests should be made with charges of equal well rounded depths and weights. Blasting of such charges is referred to as 2 4 3 2 “standard blasting”. For investigations into the liquefaction 7 Sand, fine, 00 1 4 10-12 8 43 of a sand stratum, to 10 metres deep, a charge of ammonite poorly rounded 2 2 has been chosen as “s'andard”. A weight of 5 kg of explosive (above water 3 3 and a depth of 4-5 metres of overburden will ensure that sluicing) there is no expulsion of soil during blasting. 8 — 46 10 1 24 20 Within a few seconds of firing a standard charge in a loose (under-water 2 13 saturated sandy soil, gas expulsion and surface settlem ent sluicing) 3 15 (Fig. 7), will be followed by squeezing of water through the 9 Sand, medium 37-38 0-3 1 7 3-10 2 7 pores. In some cases the water escapes as a concentrated stream 10 —— 0-3 1 0 0 or fountain, which in a loose fine sand may last as long as 110 One advantage of standard blasting is the reliability of the [2] F lorin , V. A. (1951). On the problem of liquefaction of results obtained, particularly those associated with the estima­ pure, saturated fine sands, Hidrotechnicheskoe Stroitel'stvo tion of a large area of soil subjected to blasting effects. Other (Water Power Engineering), No. 7. advantages are simplicity and low cost. This method m ay be used by geological survey expeditions as well as on construc­ [3] — (1952). The phenomena of liquefaction and methods of tion sites without the need for special supplementary equip­ compaction of loose saturated sandy foundations, Izvestia m ent. (Proceedings) o f the U.S.S.R. Academy o f Sciences, No. 6. All measures against liquefaction of soils may be divided [4] I vanov , P. L. (1956). Peculiarities of liquefaction of satu­ into two main categories : measures preventing liquefaction rated sandy soils influenced by drainage and surcharge, (sand compaction, loading, etc.) and measures for reducing Izvestia (Proceedings) o f the U.S.S.R. Academy o f Sciences, the decreasing harmful effects of liquefaction (drainages No. 8. sheet-pile bulkheads, draining surcharges, etc.). [5] — S ider as sh ., A. (1957). Laboratory investigations in the References phenomenon of liquefaction of saturated sandy soils [1] T erzaghi , K. and P eck , R. (1948). Soil Mechanics in acted on by impact and vibration influences, Proceedings Engineering Practice, New-York, London. of the Kaunas Polytechnic Institute, vol. VII, 1957.

Ill